Pandas: Selecting rows to create new dataframe using the index

Asked Nov 18 '17 at 18:41

Active Nov 18 '17 at 18:46

Viewed 1,708 times

Simple question. I have a dataframe that looks like this (call if df1):

Country 2016 2017 2018
USA 1 2 3 
CAN 4 5 6
MEX 7 5 6
ARG 9 3 5 
GBR 5 8 2

I have a variable called countries which looks like this: countries=['USA','MEX','GBR']. How can I make a new dataframe (call it df2) which selects only the rows mentioned in countries?

I can select rows manually, or several of them if they are in a block (say USA to MEX), but I don't know how to do it if the order is random. A loop could do it, (see below) but that only saves the last country in the loop!

for country in countries:
    df2 = df1.loc[country:country,:].copy()

PS. I'm sure this is a duplicate! But I've spent a good hour searching for the answer with no success.

asked Nov 18 '17 at 18:41

Agrippa

1

`df2 = df1.loc[countries]` should works – jezrael Nov 18 '17 at 18:42
1

If Country is a column, not the name of the index you need `df1[df1['Country'].isin(countries)]` – ayhan Nov 18 '17 at 18:44
OMG - I knew it was simple. So simple that no one has probably asked it before. – Agrippa Nov 18 '17 at 18:44
Country is the index. I just tried @jezrael solution and it worked. Feel like a bit of a newb! (well I am for python) – Agrippa Nov 18 '17 at 18:45
@ayhan - csn you check if is dupe for this? Because now it is marked bad... – jezrael Nov 18 '17 at 18:46
@jezrael I've reopened it. – ayhan Nov 18 '17 at 18:46
Hmmm, it is dupe - https://stackoverflow.com/q/19155718/2901002 – jezrael Nov 18 '17 at 18:48
@ahyan yep nice! works for index too with the alternative: `df2 = df1[df1.index.isin(countries)]` – Anton vBR Nov 18 '17 at 18:54

Pandas: Selecting rows to create new dataframe using the index

0 Answers0